Wing defects of shlp mutant flies and rescue of the mutant phenotype

Since shlp mutant flies were viable and fertile, we next checked whether they showed some milder developmental defects. We observed an increased number of flies homozygous for shlp¹³¹ that did not unfold their wings compared to wildtype control flies.

29.1 % of shlp¹³¹ mutant flies (n = 385) had wings that were only partially expanded or completely unexpanded. Wildtype flies did not show this defect at all (n = 241) and shlp^GE24395 flies, which carry the P-element insertion that was used to generate the mutant, showed this phenotype with a very low penetrance (0.4 %, n = 262).

In order to show that this defect is caused by the loss of shlp gene function only, we rescued it by ectopically expressing shlp in the mutant background. For this purpose, we introduced pUASt-shlp on the second chromosome in the shlp¹³¹ mutant background

shlp52 shlp52 /shlp131

A A‘ A‘‘

B B‘ B‘‘

C C‘ C‘‘

Bazooka Miranda Merge + DAPI

wildtype

41 (pUASt-shlp; shlp¹³¹) and crossed it to arm::GAL4, which itself had been crossed into the shlp¹³¹ mutant background and determined the number of flies with normally unfolded wings (Fig. A.2-18).

Figure A.2-18: Rescue of shlp mutant. Expression of untagged Shlp in the shlp¹³¹ mutant background (arm::GAL4/pUASt-shlp;shlp¹³¹) results in a complete rescue of the wing unfolding defects. If only pUASt-shlp without arm::GAL4 is present (pUASt-shlp;shlp¹³¹) no rescue was observed. The experiment was repeated three times. The experiments were performed at 25°C.

Expression of an untagged Shlp construct under the control of armadillo::GAL4 resulted in an almost complete rescue (99.5 % expanded wings) of the wing phenotype (Fig. A.2-17, arm::GAL4/pUASt-shlp; shlp¹³¹). If only pUASt-shlp was present without transactivator, we observed no rescue (9.8 % partially expanded wings; 30.7 % unexpanded wings).

The wing unfolding phenotype we observed was variable since some flies had partially unexpanded wings (Fig. A.2-19 B, B’), while others had completely unexpanded wings (Fig. A.2-19 C, C’). Additionally, we observed other defects to the wing unfolding phenotype. shlp¹³¹ mutant flies had postscutellar bristles that were frequently crossed (Fig.

A.2-19 B’, C’, white arrow), while the bristles of control animals had an almost parallel orientation (Fig. A.2-19 A’, white arrow). Pronounced crossing of postscutellar bristles has also been described for loss of function mutants of the α subunit of Bursicon (Bursicon α) and could indicate a failure in thoracic expansion (Dewey et al., 2004). While the cuticle of control animals is smooth and shiny (Fig. A.2-19 A’), the cuticle of shlp¹³¹ animals often appeared soft and blunt (Fig. A.2-19 B’, C’), resembling the cuticle of freshly hatched flies. The thorax and notum of shlp¹³¹ mutant flies were often darkened (Fig. A.2-19 C’, white asterisk). Animals that showed the blunt cuticle phenotype often had wide clefts between the abdominal segments (Fig. A.2-19 C’, white arrowhead) which sometimes occurred together with defects in abdominal segmentation (Fig. A.2-19 C, C’, black

42

arrows).

Figure A.2-19: shlp¹³¹ flies display defects in wing unfolding and cuticle tanning. (A and A’) white female with unfolded wings and normally tanned cuticle. Postscutellar bristles are oriented in parallel (white arrow). (B and B’) shlp¹³¹ female with partially expanded wings. Postscutellar bristles are crossed (white arrow). (C and C’) shlp¹³¹ male with unexpanded wings. Postscutellar bristles are crossed (white arrow). Cuticle on thorax and notum is darkened (white asterisk). Wide clefts between abdominal segments are obvious (white arrowhead). Thoracic cuticle of shlp¹³¹ animals (B’, C’) appears blunt and soft compared to cuticle of control animals (A’). Some shlp¹³¹ animals showed segmentation defects in the abdomen (C, C’, black arrows). Scale bar in A-C is 1000 µm;

500 µm in A’-C’.

We searched the literature for mutants that cause similar phenotypes and found one candidate pathway that leads to similar wing and cuticle defects when disturbed. Loss of function in Bursicon signaling causes defective wing expansion and cuticle tanning (Fraenkel and Hsiao, 1962; Baker and Truman, 2002; Dewey et al., 2004; Peabody et al., 2008). Also pronounced crossing of postscutellar bristles has been described for Bursicon mutants (Dewey et al., 2004).

We therefore checked if there is any genetic interaction between shlp¹³¹ and the gene

A A‘

B B‘

C C‘

whiteshlp131

43 rickets (rk), which encodes for the supposed Bursicon receptor Drosophila Leu-rich repeats-containing G-protein- coupled receptor 2 (DLGR2) (Baker and Truman, 2002).

Flies homozygous mutant for rk are viable and exhibits wing unfolding and cuticle tanning defects similar to what we observed for shlp loss of function. If both genes act in the same pathway and one introduces one mutant copy of rk into the shlp¹³¹ mutant background this will enhance the wing expansion defect even though animals with only one mutant copy of rk, do not exhibit wing expansion defects. Animals heterozygous for a null allele of rk and homozygous for shlp¹³¹ (rk⁴/+; shlp¹³¹) showed the low penetrance wing unfolding defect observed in shlp¹³¹ mutant flies. Surprisingly, animals homozygous for rk⁴ and shlp¹³¹ (rk⁴; shlp¹³¹) exhibited defects that indicated an enhancement of the shlp¹³¹ mutant phenotype (Fig. A.2-20 D and D’). As mentioned above, shlp¹³¹ mutant flies show a wing unfolding defect with low penetrance that was often connected to defects in cuticle morphology and intensive darkening especially of the thorax and notum (Fig. A.2-20 B and B’). Flies homozygous for rk⁴ did not expand their wings and their cuticle tanning was defective.

Although their cuticle darkened similar to white control flies (Fig. A.2-20 C-C’), the timing of this process was delayed up to several hours (Baker and Truman, 2002). rk⁴ mutants also displayed extensive crossing of postscutellar bristles indicating a defect in thorax expansion. Similar to shlp¹³¹ mutants, their cuticle often appeared blunt and soft (Figure A.2-20 C’). Flies homozygous for both rk⁴ and shlp¹³¹ displayed a complete block in wing expansion like in rk⁴ mutant flies (Fig. A.2-20 D). Similar to the shlp¹³¹ or rk⁴ homozygous flies alone postscutellar bristles of these flies showed pronounced crossing and their cuticle appeared blunt and soft (Fig. A.2-20 D’, white arrow). Unlike the rk⁴ homozygous flies, that never showed extensive darkening of the cuticle, and unlike the shlp¹³¹ homozygous flies, that showed extensive darkening of the cuticle only with moderate penetrance, the rk⁴; shlp¹³¹ flies showed this phenotype with high penetrance (Fig. A.2-20 D’, white asterisk). Therefore the rk⁴; shlp¹³¹ double mutant displays defects that neither the rk⁴ nor the shlp¹³¹ single mutant shows. This additive effect indicates that the two genes rather act in parallel pathways than in the same pathway.

44

Figure A.2-20: The rk⁴ allele enhances the shlp¹³¹ phenotype. (A and A’) white male with wildtype wings and cuticle tanning. Cuticle appears smooth and shiny. (B and B’) shlp¹³¹ male with partially expanded wings. Thorax shows abnormal darkening (white asteisk) and postscutellar bristles display prominent crossing (white arrow). (C and C’) rk⁴ male with unfolded wings and pronounced crossing of postscutellar bristles (white arrow). Cuticle appears blunt. Cuticle of the thorax is pigmented normally (white asterisk). (D and D’) rk⁴; shlp¹³¹ male with unfolded wings, abnormal darkening of the thorax (white asterisk). Postscutellar bristles show prominent crossing (white arrow). Cuticle appears blunt. Animals often have wide clefts between abdominal segments (white arrowhead). Scale bar in A, B, C and D is 1000 µm; 500 µm in A’, B’, C’ and D’.

Bursicon signaling induces cuticle hardening and tanning and regulates wing expansion.

Following wing expansion it is supposed to promote programmed cell death of epidermal cells of the wing (Kimura et al., 2004). Since shlp¹³¹ mutant flies displayed wing

whiteshlp131 rk4 ;shlp131 rk4

A A‘

B B‘

C C‘

D D‘

45 expansion defects similar to mutants that affect Bursicon signalling, we checked whether shlp¹³¹ mutants had any defects in programmed cell death in the wing epithelium.

Therefore, we collected freshly hatched flies and allowed them to develop for 30-60 min.

After this period wings were dissected, fixed and subsequently stained with the DNA marker DAPI to visualize the morphology of the cell nuclei. During this 30-60 min time period most wildtype control flies expanded their wings (Fig. A.2-21 A, A’). DAPI staining revealed that most of the wing cells had disappeared from the wing while chromatin of the remaining cells was fragmented, indicating the onset of cell death (Fig A.2-21 A’, white arrows). In a subset of wing epithelial cells nuclei were still intact (Fig. A.2-21 A’, white arrowheads). These cells represent the cells of the wing that are associated with veins.

DAPI staining on shlp¹³¹ mutant flies that expanded wings normally revealed similar results as in wildtype animals (data not shown). DAPI staining on unexpanded wings of shlp¹³¹ mutant flies revealed an increased number of viable cells (Fig. A.2-21 B, B’). After 30-60 min wing epithelial cells were intact as shown by DAPI staining and nuclei did not display chromatin fragmentation as seen in wildtype wings (Fig. A.2-21 A’). Since it has been proposed that dying cells are cleared from the wing within 3 h after wing spreading (Togel et al., 2008) we also analyzed the wing epithelium 5 h 30 min - 6 h after hatching (Fig. A.2-21 C, C’ and D, D’). At this time point the wing blade should be free of cells except for the cells of the wing vein. While the wing of control flies was free of living cells except for the cells of the veins (Fig 4.2-21 C and C’, white arrowheads), cells in unexpanded wings of shlp¹³¹ mutant flies were still alive as indicated by DAPI staining (Fig. A.2-21 D and D’). We also analyzed the wings of shlp¹³¹ mutant flies 24 h and 48 h after hatching (data not shown). While 24 h after hatching living cells were still present, no living cells could be observed after 48 h (data not shown). These results indicate that programmed cell death of wing epithelial cells is severely delayed in shlp mutant flies.

46

Figure A.2-21: Apoptosis of epithelial cells is delayed in wings of shlp¹³¹ mutant flies. (A and A’) Wing of a wildtype fly 30 - 60 min after hatching. Only nuclei in wing veins are intact (A’, white arrowheads). Nuclear staining in most other areas of the wing is absent or appears dispersed and fragmented (A’, white arrows). (B and B’) Unexpanded wing of shlp¹³¹ fly 30 – 60 min after hatching. DAPI staining reveals that the morphology of cell nuclei is intact. (C and C’) In a wing of a wildtype fly nuclear staining is only present in the wing veins, indicating loss of all other cells. (D and D’) Nuclei are still intact in a wing of a shlp¹³¹ mutant fly 5 h 30 min – 6 h after hatching. Area marked with a rectangle in A - D is magnified in A’ - D’. All images are merges of DAPI and DIC channel. Scale bar is100 µm in A - D and 50 µm in A’ - D’.

shlp131 shlp131 wildtype wildtype

A A‘

B B‘

C C‘

D D‘

30 - 60 min 5 h 30 min - 6 h

47 The defects of shlp¹³¹ mutant flies in wing unfolding and cell death of wing epithelial cells show similarities to the defects in flies with disrupted Bursicon signalling. A major difference is the low penetrance of the defects of shlp¹³¹ mutant flies compared to the highly penetrant phenotypes seen for example in rk⁴ mutants, where wing expansion is completely blocked. One explanation for these observations could be lowered Bursicon levels in the shlp¹³¹ mutant. Previously, it has been reported that Bursicon α is expressed in pairs of neurons in the subesophageal, thoracic and abdominal neuromeres of the ventral nerve cord (Dewey et al., 2004; Peabody et al., 2008; Zhao et al., 2008). To check whether the expression pattern of the heterodimeric neurohormone Bursicon is changed, we stained central nervous systems (CNS) of wandering third instar larvae with an antibody directed against Bursicon α. CNS of control wildtype larvae (A.2-22 A and A’) stained with the Bursicon α antibody showed the previously reported pattern. The pattern of Bursicon α immunoreactivity on CNS of shlp¹³¹ mutant larvae differed from the wildtype pattern.

While in wildtype CNS Bursicon α positive neurons were present in the thoracic neuromeres (Fig. A.2-22 A, white arrow), these neurons were often absent in shlp¹³¹ CNS (Fig. A.2-22 B, white arrow). We quantified the number of Bursicon α positive neurons in the CNS of wildtype and shlp¹³¹ wandering third instar larvae (Fig. A.2-22 C). In contrast to an average number of 33 neurons in wildtype CNS (n = 16), in shlp¹³¹ CNS we detected only 23 neurons on average (n = 15). These experiments demonstrate that the number of Bursicon α positive neurons is reduced in CNS of shlp¹³¹ mutant third instar larvae.

48

´

Figure A.2-22: The number of Bursicon α positive neurons is reduced in the ventral nerve cord of shlp¹³¹ L3 larvae. (A and A’) In wildtype third instar larvae, Bursicon α is expressed in pairs of neurons in the subesophageal, thoracic and abdominal neuromeres of the ventral nerve cord. (B and B’) In the CNS of shlp¹³¹ larvae, Bursicon α is often not expressed in the thoracic neuromeres (white arrow, compare with white arrow in A). Green = Bursicon α, turquoise = DAPI. A, A’, B and B’ are maximum intensity projections of single confocal planes. Scale bar is 100 µm. (C) Quantification of the number of Bursicon α positive neurons reveals shlp¹³¹ larvae have on average less Bursicon α positive neurons (23±5, n=15) than wildtype larvae (33±6, n=16).

0 5 10 15 20 25 30 35 40 45

1 2

Average number of Bursicon αpositive neurons

1.wildtype 2. shlp¹³¹

C

C

Bursicon α Bursicon α DAPI

wildtype shlp131

A A‘

B B‘

49 A.2.5.3 Overexpression of Shlp-eGFP in flies mimics the defects seen in shlp¹³¹ mutant flies

Embryos overexpressing eGFP developed normally and the localization of Shlp-eGFP in the embryo was similar to untagged Shlp when overexpressed (see A.2.2) Surprisingly, when we overexpressed Shlp-eGFP with the strong ubiquitous GAL4 line tub::GAL4 we observed wing and cuticle phenotypes that were comparable to those seen in the shlp¹³¹ mutant (Fig. A.2-23 A and A’). Interestingly, all flies that overexpressed Shlp-eGFP under the control of tub::GAL4 showed the wing unfolding and cuticle defects (Fig.

4.2-23 B and B’). This effect seemed to be dosage dependent, since we only observed this fully penetrant phenotype when using tub::GAL4, while using other ubiquitous transactivator lines like actin::GAL4 or da::GAL4 resulted only in a moderate number of flies with wing defects (data not shown).

Figure A.2-23: Flies overexpressing Shlp-eGFP show defects in wing unfolding and cuticle tanning. (A and A’) shlp¹³¹ male with unexpanded wings. (A’) Postscutellar bristles are crossed (white arrow). Cuticle on thorax and notum is darkened (white asterisk) and there are wide clefts between abdominal segments (white arrowhead). (B and B’) male fly expressing Shlp-eGFP under the control of Tub::GAL4. Wings are unexpanded, postscutellar bristles are crossed (white arrow), cuticle on notum is darkened (white asterisk) and there are wide clefts between abdominal segments (white arrowhead). Cuticle appears blunt. Scale bar in A, B is 1000 µm; 500 µm in A*

and B*.

To check whether clearance of epithelial cells is defective just as in shlp¹³¹ mutant flies, we fixed and stained wings of flies overexpressing Shlp-eGFP 30 – 60 min and 5:30 – 6 h after hatching as described above. After 30 – 60 min, identical to wings of shlp¹³¹ mutant flies that did not unfold their wings, wings of Shlp-eGFP overexpressing flies were positive for

A A‘

B B‘

Tub::GAL4>pUASt ::Shlp-GFPshlp131

50

DAPI staining and no sign of chromatin fragmentation was visible (Fig. A.2-24 A and A’).

Even after 5:30 – 6 h nuclei of wing epithelial cells, as revealed by DAPI staining, appeared to be intact (Fig. A.2-24 B and B’).

Figure A.2-24: Apoptosis of epithelial cells is delayed in the wing of flies that overexpress Shlp-eGFP. (A and A’) Unexpanded wing of a fly overexpressing Shlp-eGFP under the control of tub::GAL4 30 – 60 min after hatching. DAPI staining reveals that wing cells are intact. (B and B’) Cells are still intact after 5:30 – 6 h after hatching. Area marked with a rectangle in A and B is magnified in A’ and B’. All images are merges of DAPI and DIC channel. Scale bars are 100 µm in A and B and 50 µm in A’ and B’.

A.2.5.4 Functional analysis of Shlp protein domains

We have shown that the wing unfolding defects displayed by flies homozygous for shlp¹³¹ can be rescued by expressing an untagged form of Shlp (see chapter A.2.5.2).

Overexpression of carboxy-terminally eGFP-tagged Shlp protein in the fly causes similar defects as shlp¹³¹, indicating a potential dominant negative function of this protein. To analyze which domains are needed to rescue the wing expansion defects of shlp¹³¹ flies, we generated a series of Shlp deletion constructs with an amino-terminally attached eGFP and generated transgenic flies (Fig. A.2-25). The transgenes inserted on the second chromosome were selected and crossed into the shlp¹³¹ background analogous as described in A.2.5.2. Like the wildtype protein, the deletion constructs were scored for rescue of the wing expansion and cuticle tanning defects seen in the shlp¹³¹ mutant. Final results of the rescue experiments are still in progress.

tub::GAL4>pUASt::Shlp-GFP

A A‘

B B‘

30 - 60 min 5:30 – 6 h

51

Figure A.2-25: Schematic representation of Shlp constructs used to generate transgenic flies.

Shlp-eGFP is wildtype Shlp with an eGFP tag at the carboxy terminus. eGFP-Shlp is wildtype Shlp with an amino terminal eGFP tag. eGFP-Shlp∆589-596 lacks the last eight carboxy terminal aa of Shlp; eGFP-Shlp∆C lacks the last 22 carboxy terminal aa; eGFP-Shlp∆TM+C lacks the transmembrane domain and the intracellular domain; eGFP-Shlp∆exDom lacks the extracellular domain of Shlp. For all amino-terminally eGFP tagged proteins, the Shlp signal peptide was placed in front of the eGFP tag to achieve membrane targeting of the proteins.

A.2.6 Infection and survival experiments to test a potential role of Shlp in immunity

In cooperation with the group of Bruno Lemaitre survival experiments after bacterial infection were conducted to test if Shlp has an influence on the humoral immune reaction.

We conducted these experiments because TIP, the mammalian homologue of Shlp, has a modulatory role in adaptive immune responses (Fiscella et al., 2003). Although Drosophila has no adaptive immune response, many of the mechanisms governing innate immunity are conserved between vertebrates and invertebrates and therefore one could speculate that Shlp, since it is not required for viability, might have a function during the innate immune

Shlp-eGFP

Shlp

eGFP-Shlp

eGFP-Shlp∆589-596

eGFP-Shlp∆C

eGFP-Shlp∆TM+C

eGFP-Shlp∆exDom

52

response. To test the role of Shlp in humoral immunity, shlp¹³¹ mutant flies were challenged by infection with the Gram-negative bacterium Erwinia carotovora (E.

carotovora) (Fig. A.2-26 A) and the Gram-positve bacterium Enterococcus faecalis (E.

faecalis) (Fig. A.2-26 B). It is known that infection of flies with E. carotovora activates the Toll pathway, while infection with E. faecalis activates the immune deficiency (Imd) pathway (Buchon et al., 2009). In response to infection with E. carotovora shlp¹³¹ flies survived comparable to wildtype or shlp^GE24395 flies (Fig. A.2-26 A). As a control a relish (rel) loss-of-function mutant (rel^E20) was included in the experiment. In the rel^E20 mutant, Imd pathway mediated induction of antimicrobial peptides is impaired (Janeway and Medzhitov, 2002). Rel^E20 flies in contrast to wildtype, shlp^GE24395 and shlp¹³¹ flies rapidly succumbed infection with E. corotovora. shlp¹³¹ mutant flies survived infection with E.

faecalis comparably well as wildtype flies, while null mutants for spätzle (spl^rm7), died quickly after infection (Fig. A.2-26 B). spätzle encodes a ligand that is required to induce the Toll pathway and spl loss-of-function leads to loss of Toll pathway induction (Ferrandon et al., 2007) and therefore flies mutant for spl are susceptible to infection with Gram-positive bacteria. The increased survival of shlp^GE24395 flies in comparison to wildtype flies was not further considered here and may be a matter of statistical variation.

We conclude from these experiments that shlp is not directly involved in the humoral immune reaction in response to bacterial infection.

53

Figure A.2-26: shlp¹³¹ mutant flies do not show an increased susceptibility to infection with Gram-negative (A) or Gram-positive (B) bacteria. The survival rates (%) of shlp¹³¹ flies infected with the Gram-negative bacterium Erwinia carotovora (A) or the Gram-positive bacterium Enterococcus faecalis (B) were compared with shlp^GE24395 flies, wildtype (wt) flies and flies mutant wither in the Imd pathway (rel^E20) or the Toll pathway (spz^rm7). Flies were infected by septic injury with a needle dipped in a concentrated bacterial pellet.

A.2.7 Biochemical characterization of Shlp protein A.2.6.1 Glycosylation of Shlp

The mammalian homologue of Shlp, TIP, has been described to harbor twelve potential N-linked glycosylation sites. When performing Western Blot analysis with an antibody directed against TIP on the purified extracellular domain of TIP which had been expressed in 293T mammalian cell culture cells, the protein ran at a position higher than the predicted molecular weight (Fiscella et al., 2003). Therefore, we analyzed the Shlp amino acid sequence for the presence of N-linked glycosylation sites using the program NetNGlyc 1.0 (http://www.cbs.dtu.dk/services/NetNGlyc/). Six asparagine residues at the amino acid positions 22, 129, 167, 187, 229 and 379 are predicted to be N-glycosylated according to

0

0 14 24 38 62 70 86 94 110117138164165

survival ((%)

54

250

this algorithm, out of which glycosylation of residues 22, 129 and 187 is most likely. To test if Shlp is N-glycosylated, we performed a deglycosylation assay. Protein lysates were prepared form embryos expressing Shlp-eGFP ubiquitously under the control of da::GAL4.

Then Shlp-eGFP was precipitated using GFP-Trap^® as described in Material and Methods C.6.2and treated with PNGase F. PNGase F, also known as N-Glycosidase F, is an amidase that cleaves between the innermost N-acetylglucosamin and asparagine residues of high mannose, hybrid and complex oligosaccharides from N-linked glycoproteins (Maley et al., 1989). Shlp-eGFP has a predicted molecular weight of 94 kDa. But detection of Shlp-eGFP in Western Blot shows a molecular weight of approximately 115 kDa (Fig. A.2-26, second lane from left). Shlp-eGFP treated with PNGase F ran at a molecular weight of

Then Shlp-eGFP was precipitated using GFP-Trap^® as described in Material and Methods C.6.2and treated with PNGase F. PNGase F, also known as N-Glycosidase F, is an amidase that cleaves between the innermost N-acetylglucosamin and asparagine residues of high mannose, hybrid and complex oligosaccharides from N-linked glycoproteins (Maley et al., 1989). Shlp-eGFP has a predicted molecular weight of 94 kDa. But detection of Shlp-eGFP in Western Blot shows a molecular weight of approximately 115 kDa (Fig. A.2-26, second lane from left). Shlp-eGFP treated with PNGase F ran at a molecular weight of

Im Dokument Functional characterization of the gene schlappohr (CG7739) during Drosophila development (Seite 44-0)